I heard the pollster Mark Penn say once that most people like numbers – they just don’t like math. It’s a good thing to keep in mind whenever you have to make a data-heavy presentation.
Data analysis can be daunting for novices, even when it doesn’t go beyond arithmetic. Data discovery tools, which arose to give business users more freedom to analyze data than their business intelligence platforms tools allowed, appeal to this desire to make analytics easier and less scary for businesses. BI tools are heavy, expensive and hard to use. Data discovery was to be the opposite.
When you compare the two kinds of products, there’s no quesiton that data discovery tools are simpler and lighter. But do users feel they’ve gotten a radically simpler way for everyone to be able to analyze data?
We read 80 reviews of data discovery tools inlcuding Tableau, Microsoft Power BI, QlikView, Looker, IBM Watson and more — all written by UserMuse members. The results, I’m a bit sorry to say, were what we expected.
“Easy to Use” Can Still Be Confusing
Data discovery tools face a flexibility-usability tradeoff: Should you allow users to ask any possible question and risk they make mistakes, get stuck, or break something? Or do you restrict functionality and data access to prevent those kinds of user errors?
On a 1-5 scale with 5 being the highest, the category overall scored a 4.15 from UserMuse members. Tableau had the highest score among companies with more than ten reviews (4.33), while Looker had the lowest in our sample (3.83). Users complimented the various tools’ visualization, sharing, and data import features as reasons for high satisfaction.
When asked what about the tools they use failed to meet expectations, 56% of respondents cited that the tools were too difficult to set up and use. A sample of responses:
Understanding exactly how Tableau’s internal built-in calculations were working could be incredibly frustrating. Some seemingly simple built-in calculations often returned confusing results that could only be found in detailed review. – Tableau User
The UX is cumbersome and requires a fair amount of training for teams – it is not always easy to self-navigate. – Looker User
You still need to have the fundamentals down. This isn’t a detriment to the product, but the product is not magic. You still need to know how to structure data. – Qlik User
If data discovery products were cigarettes, “You still need to know how to structure data” would be plastered all over the pack. Still, I read these critiques largely as observations about working with enterprise data in general and less about the tools specifically. Here’s why.
Why It’s So Hard to “Democratize” Data Analysis
Being an effective data analyst requires three things:
- The business acumen and creativity to ask important questions
- Thorough understanding of the underlying data
- Mastery of the techniques and tools needed to analyze the data
An effective data discovery product helps people who have all three answer business questions. The problem we see is that many buyers seem to expect (with ample nudging from the marketing they see) that the tool magically does #2 and #3 for them. But a tool can’t give you either of those things, and without either of them you can’t get much value from the tool.
Not everyone can be a data analyst, just as not everyone can be a marketer, salesperson, or manager. This category is full of excellent products – but also plenty of hype and marketing.